hosts stuck in refresh when inotify limit is reached
stbuehler opened this issue · comments
Stefan Bühler commented
When the system inotify limit is reached hosts are stuck in the refresh state.
Some node
based applications are known to trigger the inotify limit (see microsoft/vscode#7950).
I see no real need for apt-dater to rely on inotify. The internal event system ("refresh finished") should not rely on it at all.
For the record, you see something like this in strace for one thread over and over:
inotify_add_watch(7, "/home/XXX/.cache/apt-dater/stats", IN_MODIFY|IN_ATTRIB|IN_CLOSE_WRITE|IN_MOVED_FROM|IN_MOVED_TO|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF|IN_UNMOUNT|IN_ONLYDIR) = -1 ENOSPC (No space left on device)
inotify_add_watch(7, "/home/XXX/.cache/apt-dater/tmux", IN_MODIFY|IN_ATTRIB|IN_CLOSE_WRITE|IN_MOVED_FROM|IN_MOVED_TO|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF|IN_UNMOUNT|IN_ONLYDIR) = -1 ENOSPC (No space left on device)
[...]
inotify_add_watch(7, "/home/XXX/.cache/apt-dater/stats", IN_MODIFY|IN_ATTRIB|IN_CLOSE_WRITE|IN_MOVED_FROM|IN_MOVED_TO|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF|IN_UNMOUNT|IN_ONLYDIR) = -1 ENOSPC (No space left on device)
inotify_add_watch(7, "/home/XXX/.cache/apt-dater/tmux", IN_MODIFY|IN_ATTRIB|IN_CLOSE_WRITE|IN_MOVED_FROM|IN_MOVED_TO|IN_CREATE|IN_DELETE|IN_DELETE_SELF|IN_MOVE_SELF|IN_UNMOUNT|IN_ONLYDIR) = -1 ENOSPC (No space left on device)
poll([{fd=3, events=POLLIN}, {fd=7, events=POLLIN}], 2, 3997) = 0 (Timeout)
Thomas Liske commented
Since you could run multiple apt-dater instances in parallel it relys on inotify to update the host status. By removing the usage of inotify it would break apt-dater when running multiple instances (of different users).